截断情况下回归函数的经验似然推断

Empirical Likelihood for Regression Function with Censored Data

讨论在数据截断情况下回归函数的经验似然比置信区间的统计推断,其中,响应变量的估计采用classk估计,响应随机变量和删失随机变量的分布函数用其对应的K-M函数替代,得到调整的对数似然比统计量渐近x12分布的结果.

In this paper, the (log)empirical likelihood confidence intervals for regression function with censored data are investigated.The response variables are replaced by the class k estimators,and the distribution funtions of the response variables and the random censoring variables are substituted by the K-M estimators,respectively. It is shown that the abjusted (log)empirical likelihooh follows a chi -square distribution.